APPLICATION OF DATA MINING AND COLLABORATIVE FILTERING BASED ON STUDENT INFORMATION EXTRACTION IN THE CONSTRUCTION OF RECOMMENDATION SYSTEMS IN UNIVERSITY LIBRARIES

University libraries are important knowledge bases for teachers and students to acquire knowledge. It is necessary to study how to process big data, extract features, and provide high-quality personalized services to users in order to improve the efficiency of library book recommendations. This study proposes a recommendation algorithm based on data mining and collaborative filtering algorithms for book recommendation systems. The data are processed and selected using data extraction and decision tree methods to obtain useful data information. Clustering and improved collaborative filtering algorithms based on users are used to process the data and obtain the similarity between users, which serves as the basis for selecting books. Relevant experiments were designed for validation in the experiment. These experiments confirm that as the number of book recommendations increases, the maximum difference during the growth process is 16.17% for 10 books. For 15 books, the accuracies of the traditional algorithm and the improved algorithm were 75.91% and 85.79%, respectively, with a difference of 9.88%. Through a questionnaire survey, the overall average satisfaction rate of teachers and students is 62.12%. Among teachers, the highest satisfaction rate is 68.25%, while the lowest satisfaction rate is 48.52%. Among students, the highest and lowest satisfaction rates are 78.35% and 51.34%, respectively. Therefore, the application of this recommendation algorithm in book recommendation systems has a good promoting effect on the management and personalized recommendation of university libraries. It has certain significance in providing book recommending services and making decisions to make library collections' layout optimized.